Platinum Metals Rev., 1989, 33, (2), 64
Properties of Binary Rhodium Alloys
Rhodium has a higher melting point, greater specific strength and better oxidation resistance than most of the refractory and platinum group metals but, to date, neither rhodium nor its alloys have found wide application as structural materials. During a recent study of rhodium alloys a vast amount of data was collected, and a selection of this information is recorded here. An investigation of the mechanical properties of some of these alloys indicates that a number of them could be useful for specific applications that take place under arduous conditions.
Crucibles made of rhodium have found limited use, specifically during the growth of calcium tungstate (CaWO4) and barium metatitanate (BaTiO3) single crystals, being selected for this application because rhodium maintains its shape better than zirconia grain stabilised (ZGS) platinum alloys at the operating temperature range of 1610 to 1650°C (1). Additionally a number of rhodium alloys have been considered as structural materials for use in the glass industry, but they did not find practical application because rhodium dissolved from the fabricated product and discoloured the glass. Now, however, advances in metal melting technology have made it possible to produce a large number of new rhodium alloys, the development and properties of which are summarised here.
At the start of this investigation a survey of published information revealed that most elements have a solid solubility of 2 to 15 atomic per cent in rhodium, as shown by their phase diagrams, but only limited information about the mechanical properties of rhodium alloys was discovered (2–7). For this reason a series of dilute rhodium alloys was prepared by argon arc melting pure elements. These alloys are listed in Table I. The concentration of the additional elements was 1 weight per cent, which for most alloys is below the solid solubility limit and thus prevents precipitation of a second phase. The cast alloys were then fabricated into sheets, and their stress rupture properties determined at loadings of 110 to 345 bar, at temperatures of 1200, 1300 and 1400°C. As a result of an initial study, further alloys were then prepared to enable the effect of the concentration of refractory group metals and indium upon the creep life of these rhodium alloys to be determined.
As-cast hardness values for binary rhodium alloys containing 1 weight per cent of the alloying element are given in Table I. Most elements produced a small hardness increase of 20 to 70 Hv, compared to the figure for pure rhodium. However, the addition of boron, selenium or cerium resulted in a significant increase in hardness, to 665, 446 and 304 Hv, respectively; but of these only the cerium alloy was workable. Later the amount of the alloying element was reduced to 0.5 weight per cent for the hafnium, gadolinium and dysprosium alloys, and to 0.1 weight per cent for boron-, sulphur-and phosphorus-containing rhodium alloys, because higher concentrations made the alloys unworkable.
The addition of 0.5 weight per cent hafnium and 1 weight per cent tungsten, rhenium, platinum or ruthenium improved the cold formability of annealed hot-rolled rhodium sheet, as shown by the cupping test results given in Table II.
|Cupping Tests on Hot-Rolled Rhodium Binary Alloy Sheet Containing 1 Weight Per Cent of the Second Element number of turns*|
|Mg 22||Al 11||P 23|
|Ca 34||Sc 8||Ti 30||V 30||Cr 29||Mn 39||Fe 46||Co 47||Ni 52||Cu 33||Zn 74||Ge 32|
|Sr 51||Y 21||Zr 42||Nb 40||Mo 41||Ru 53||Rh 51||Pd 45||Ag 41||Cd 66||In 30||Sn 33||Sb 29|
|Ba 46||La 17||Hf** 55||Ta 43||W 71||Re 54||Os 36||Ir 49||Pt 56||Au 45||Pb 38||Bi 29|
|Ce 22||Pr 19||Nd 19||Sm 19||Gd**26||Tb 29||Dy**24||Ho 33||Er 28||Tm 26||Yb 35||Lu 33|
The principal mode of oxidation of rhodium alloys at temperatures of 1200 and 1400°C was by grain boundary diffusion and internal oxidation. Alloys which contained hafnium, niobium or zirconium developed the most protective oxide layer. The surface of the hafnium-containing alloy revealed slight pitting, and the oxide layer which had formed after 281 hours at temperature was only 0.002mm thick, compared with 0.1 to 0.3mm for the other refractory group metals. Initial calculations based on thickness measurements made of the oxide layers formed at 1200°C on alloys containing refractory group metals revealed that the rate of oxidation was controlled by the diffusion of the alloying element. The protective oxide layer which formed on zirconium alloys made them very difficult to weld during fabrication trials. Most of the alloys revealed an increased weight as the alloying element was converted into oxide.
Glass compatibility tests showed that several of the alloys listed in Table III did not colour liquid crown glass at 1200°C. Apparently the dissolution of rhodium into the liquid glass had been prevented by either the alloying element or by the formation of a protective oxide layer which prevented colouration of the liquid glass. The only alloy which did not display a substantially increased contact angle with this glass was 1 weight per cent gold-rhodium.
At both 1200 and 1400°C, the effect of most alloying elements was to increase the stress rupture life compared to that of rhodium, at a loading of 345 bar (Table IV). The alloys with the longest creep lives at 1200°C were those containing scandium, zirconium, holmium and lutetium, and at 1400°C niobium and tantalum. The creep resistance of these rhodium alloys was less than that of ZGS 10 per cent rhodium-platinum; the elongation to fracture for the rhodium alloys was a minimum of 3.9 per cent at 1200°C, and 6.6 per cent at 1400°C, compared with 0.1 per cent for ZGS 10 per cent rhodium-platinum at both these test temperatures. During creep testing rhodium alloys became very brittle due to grain boundary cavitation. This embrittlement may limit the usefulness of these rhodium alloys, as they would not be suitable for use in highly stressed components.
|Stress Rupture Properties of Binary Alloys of Rhodium, Containing 1 Weight Per Cent of the Second Element, at a Loading of 345 bar and at temperatures of 1200°C (upper) and 1400°C (lower) time to failure, in hours|
The effect of the concentration of refractory group metals and indium on creep life at 1200°C and a loading of 345 bar is shown in Figures 1 to 3. As the concentration of hafnium, zirconium, niobium, vanadium and indium was increased, a maximum stress rupture life was reached for each alloy. These maxima may have been produced by the optimum dispersion of oxide particle size and distribution. At higher concentrations the stress rupture life decreased as the oxide particle size distribution exceeded the optimum, and promoted large particles at the grain boundaries which acted as crack initiators. The alloys containing titanium, chromium and molybdenum revealed an initial peak in the stress rupture life, and then a higher secondary peak; this effect may have been caused by the precipitation of a second metallic phase. The maximum creep life of alloys containing tungsten and tantalum was less affected by concentration and may have been caused by solid solution hardening. Small additions of 0.1 per cent boron or 0.1 per cent carbon to rhodium increased the creep life from 0.6 hours to 13.3 hours and 25.1 hours, respectively, at 1200°C and a loading of 345 bar. At 1400°C, the maximum creep life of alloys containing titanium, zirconium, vanadium and molybdenum occurred with higher concentrations of the alloying elements, this data being collected in Table V.
At applied stress values of 110 and 207 bar the rhodium alloys are weaker than ZGS 10 per cent rhodium-platinum, at 1200°C, as is shown in Figure 4. At the relatively low stress of 110 bar, the most effective alloy addition was 0.1 weight per cent silicon, this alloy having a creep life of 2147.6 hours (Table VI). The premature failure of rhodium alloys at these low stress values was due to oxygen diffusion. This caused oxidation of the alloying elements, and the resultant large oxide particles acted as crack initiators; in addition the oxygen caused cavitation to occur at grain boundary triple points.
The effect of temperature upon the mechanical properties of a 0.75 per cent niobium-rhodium alloy is shown in Figure 5. Both proof stress and tensile strength decreased rapidly at 900°C, with the onset of recovery. Nevertheless this alloy was twice as strong as a cobalt based alloy from which most spinner baskets used for the manufacture of continuous fibres are generally made, and had a strength to weight ratio of 1.22 bar/g compared with 1.05 bar/g for the cobalt based alloy.
If the Lewis acid base stabilisation theory is applied for oxides (8), it can be used to explain the improvement in stress rupture lives of dilute solutions of the transition metals zirconium, hafnium and niobium binary alloys. This shows that the improvement in the creep strength of binary rhodium alloys with refractory group metals can be explained by examining the binding energy of AB compounds between these metals. Group III and IV elements such as scandium and zirconium produced the largest binding energy with rhodium, and they also produced the longest stress rupture lives.
The stress rupture results show that at high loadings some rhodium alloys are superior to ZGS 10 per cent rhodium-platinum, but at lower loadings failure occurs more rapidly because of inherent grain boundary weakness. At high loadings the creep life is dependent upon the particle size distribution and the grain size. Alloys which internally oxidised rapidly, such as scandium-rhodium, and which had a very small grain size and a fine oxide particle size distribution, produced the longest creep life. At lower stresses creep failure was dependent upon the strength of the grain boundaries and upon microslip and diffusion. Under these conditions the rhodium alloys failed more quickly than ZGS 10 per cent rhodium-platinum, due to rapid diffusion of oxygen which caused internal oxidation and produced large oxide particles which acted as crack initiators. These oxides also prevented grain boundary sliding and microslip and recovery which would have prevented cavitation at the triple points. Failure occurred rapidly in alloys which contained elements that form volatile oxides, such as tungsten, carbon and sulphur, or which develop a non-protective oxide layer. These volatile oxides either promoted void formation or failed to prevent oxygen diffusing to existing voids at the grain boundaries. It is thought that the improvement in the creep life caused by the addition of indium may have increased the grain boundary cohesion.
This study has shown that it is now possible to manufacture many binary rhodium alloys, some of which have properties that may be useful commercially. For example, it is possible to make a rhodium alloy which will not colour glass, and which can be fabricated into products with an acceptable creep life. At high stress several of these rhodium alloys have a creep life superior to that of ZGS 10 per cent rhodium-platinum.
In addition, on certain rhodium alloys it is possible to produce a protective oxide layer which prevents the metal loss normally associated with the platinum group metals.
A similar account of work on ternary and other complex rhodium alloys will appear in a subsequent issue of this journal.
- 1 G. Reinacher, Metall, 1963, 17, (7), 699
- 2 E. Savitsky,, V. Polyakova,, N. Gorina and N. Roshan, “ Physical Metallurgy of Platinum Metals ”, Mir Publishers, Moscow, 1978, pp. 180 – 316
- 3 “ Rhodium ”, International Nickel Limited, London
- 4 L. Schellenberg,, J. L. Jorda, J. Muller, J. Less-Common Met., 1985, 109, (2), 261
- 5 R. P. Elliot, “ Constitution of Binary Alloys, First Supplement ”, McGraw-Hill, New York, 1965
- 6 W. G. Moffatt, “ Handbook of Binary Phase Diagrams ”, General Electric Company, New York, 1978, Vol . 1 – 4
- 7 F. A. Shunk, “ Constitution of Binary Alloys, Second Supplement ”, McGraw-Hill, N.Y., 1969
- 8 J. K. Gibson,, L. Brewer and K. A. Gingerich, Metall. Trans., 1984, 15A, (11), 2075